2018 NATIONAL SURVEY ON DRUG USE AND HEALTH

PUBLIC USE FILE CODEBOOK

Substance Abuse and Mental Health Services Administration Center for Behavioral Health Statistics and Quality

Rockville, Maryland 20857

October 22, 2019

2018 NATIONAL SURVEY ON DRUG USE AND HEALTH

PUBLIC USE FILE CODEBOOK

Contract No. HHSS283201700002C

Project No. 0215638

Deliverable 56

For questions about this codebook and data file, please e-mail Sarra.Hedden@samhsa.hhs.gov.

Prepared for: Substance Abuse and Mental Health Services Administration, Rockville, Maryland

Prepared by: RTI International, Research Triangle Park, North Carolina

Data File Name: PUF2018_100819 October 22, 2019

Recommended Citation: Center for Behavioral Health Statistics and Quality. (2019). 2018 National Survey on Drug Use and Health Public Use File Codebook, Substance Abuse and Mental Health Services Administration, Rockville, MD

Table of Contents

Section Page

Introduction to the 2018 National Survey on Drug Use and Health ............................................. i-1 Introduction ....................................................................................................................... i-1 Overview of NSDUH ........................................................................................................ i-2

Survey Redesign in 1999 and Improvements in 2002 .......................................... i-2 Partial Questionnaire Redesign in 2015 ................................................................ i-3 Summary of Information for 2018 ........................................................................ i-3

Questionnaire Changes for the 2018 NSDUH .................................................................. i-5 Additional Historical Changes to NSDUH Variables of Note.......................................... i-6 Strengths and Limitations of NSDUH .............................................................................. i-7 Survey Methodology ......................................................................................................... i-8 Stratification and Selection of Primary, Secondary, and Tertiary Sampling Units

(Census Tracts, Census Block Groups, and Area Segments) ..................................... i-9 Selection of Dwelling Units ............................................................................................ i-10 Selection of Individuals .................................................................................................. i-11 Sample Design Variables ................................................................................................ i-12 Data Collection and Response Rates .............................................................................. i-13 Sample Weights .............................................................................................................. i-14 Organization of the Data File .......................................................................................... i-15 Usable Cases ................................................................................................................... i-16 Logical Editing................................................................................................................ i-18

Editing Procedure for Substance Use Variables Other than Prescription Drugs ............................................................................................................. i-19

Editing Procedure for Prescription Drug Variables ............................................ i-20 Standard Code Conventions ................................................................................ i-21

Statistical Imputation ...................................................................................................... i-22 Imputation Indicators .......................................................................................... i-24 Constraints and Consistency ............................................................................... i-26

Variance Estimation of Estimated Numbers of Individuals ........................................... i-26 Statistical Significance of Differences ............................................................................ i-29 Confidentiality of Data ................................................................................................... i-30 Public Use File Weight Calibration ................................................................................ i-32 Public Use File Estimates and Standard Errors .............................................................. i-33 Special Types of Analyses .............................................................................................. i-35

List of Tables

Table Page

1. Demographic Domains for the Public Use File That Should Use the AlternativeStandard Error Estimation Method for Calculating the Estimated Number ofIndividuals, Totals: 2018 ................................................................................................ i-28

2. Past Month Marijuana Prevalence and Standard Error Ratios of the Public UseFile Subsample to the Full Sample, by Seven Domains: 2018 ....................................... i-34

3. Past Year Heroin Prevalence and Standard Error Ratios of the Public Use FileSubsample to the Full Sample, by Seven Domains: 2018 .............................................. i-34

TABLE OF CONTENTS

Codebook Creation Date: 10/8/2019 ....................................................... 1

TABLE OF CONTENTS

IDENTIFICATION ....................................................................................................................................................................................................... 3SELF-ADMINISTERED SUBSTANCE USE SECTIONS .......................................................................................................................................... 4

TOBACCO ........................................................................................................................................................................................................ 4ALCOHOL ...................................................................................................................................................................................................... 17MARIJUANA .................................................................................................................................................................................................. 22COCAINE ........................................................................................................................................................................................................ 26CRACK ............................................................................................................................................................................................................ 30HEROIN .......................................................................................................................................................................................................... 34HALLUCINOGENS ........................................................................................................................................................................................ 38INHALANTS ................................................................................................................................................................................................... 53METHAMPHETAMINE ................................................................................................................................................................................. 62PAIN RELIEVERS SCREENER ..................................................................................................................................................................... 66

TRANQUILIZERS SCREENER ................................................................................................................................................................................ 67STIMULANTS SCREENER ........................................................................................................................................................................... 68SEDATIVES SCREENER ............................................................................................................................................................................... 69PAIN RELIEVERS .......................................................................................................................................................................................... 70TRANQUILIZERS .......................................................................................................................................................................................... 79STIMULANTS ................................................................................................................................................................................................ 86SEDATIVES .................................................................................................................................................................................................... 94

IMPUTED SUBSTANCE USE ................................................................................................................................................................................ 100RECENCY OF DRUG USE .......................................................................................................................................................................... 100PAST YEAR FREQUENCY OF USE ........................................................................................................................................................... 109PAST MONTH FREQUENCY OF USE ....................................................................................................................................................... 112AGE/DATE OF FIRST DRUG USE ............................................................................................................................................................. 117RECODED DRUG USE ................................................................................................................................................................................ 127

OTHER SELF-ADMINISTERED SECTIONS......................................................................................................................................................... 163SPECIAL DRUGS ......................................................................................................................................................................................... 163

RECODED SPECIAL DRUGS .............................................................................................................................................................. 175RISK/AVAILABILITY ................................................................................................................................................................................. 177

RECODED RISK/AVAILABILITY ....................................................................................................................................................... 182BLUNTS ........................................................................................................................................................................................................ 184SUBSTANCE DEPENDENCE AND ABUSE .............................................................................................................................................. 188

IMPUTED SUBSTANCE DEPENDENCE AND ABUSE ..................................................................................................................... 252RECODED SUBSTANCE DEPENDENCE AND ABUSE .................................................................................................................... 262

SPECIAL TOPICS ......................................................................................................................................................................................... 271RECODED SPECIAL TOPICS .............................................................................................................................................................. 279

PRIOR SUBSTANCE USE ........................................................................................................................................................................... 281DRUG TREATMENT ................................................................................................................................................................................... 298

RECODED DRUG TREATMENT ......................................................................................................................................................... 333HEALTH........................................................................................................................................................................................................ 349

RECODED HEALTH ............................................................................................................................................................................. 363ADULT MENTAL HEALTH SERVICE UTILIZATION ............................................................................................................................. 364

RECODED ADULT MENTAL HEALTH SERVICE UTILIZATION .................................................................................................. 381SOCIAL ENVIRONMENT ........................................................................................................................................................................... 388YOUTH EXPERIENCES .............................................................................................................................................................................. 390

RECODED YOUTH EXPERIENCES .................................................................................................................................................... 405MENTAL HEALTH ...................................................................................................................................................................................... 411

RECODED MENTAL HEALTH ............................................................................................................................................................ 420ADULT DEPRESSION ................................................................................................................................................................................. 426

RECODED ADULT DEPRESSION ....................................................................................................................................................... 444YOUTH MENTAL HEALTH SERVICE UTILIZATION ............................................................................................................................ 449

RECODED YOUTH MENTAL HEALTH SERVICE UTILIZATION .................................................................................................. 473ADOLESCENT DEPRESSION ..................................................................................................................................................................... 493

RECODED ADOLESCENT DEPRESSION .......................................................................................................................................... 512CONSUMPTION OF ALCOHOL ................................................................................................................................................................. 518

RECODED CONSUMPTION OF ALCOHOL ....................................................................................................................................... 528MARIJUANA PURCHASES ........................................................................................................................................................................ 532

INTERVIEW INFORMATION ................................................................................................................................................................................ 548DEMOGRAPHICS ................................................................................................................................................................................................... 549

DEMOGRAPHICS ........................................................................................................................................................................................ 549IMPUTED DEMOGRAPHICS ............................................................................................................................................................... 554RECODED DEMOGRAPHICS .............................................................................................................................................................. 555

EDUCATION ................................................................................................................................................................................................ 557RECODED EDUCATION ...................................................................................................................................................................... 561

EMPLOYMENT ............................................................................................................................................................................................ 562IMPUTED EMPLOYMENT................................................................................................................................................................... 568

TABLE OF CONTENTS

Codebook Creation Date: 10/8/2019 ....................................................... 2

HOUSEHOLD COMPOSITION (ROSTER) ................................................................................................................................................. 569PROXY INFORMATION.............................................................................................................................................................................. 571HEALTH INSURANCE ................................................................................................................................................................................ 572

IMPUTED HEALTH INSURANCE ....................................................................................................................................................... 577RECODED HEALTH INSURANCE ...................................................................................................................................................... 579

INCOME ........................................................................................................................................................................................................ 580IMPUTED INCOME .............................................................................................................................................................................. 581RECODED INCOME ............................................................................................................................................................................. 583

FI DEBRIEFING QUESTIONS ................................................................................................................................................................................ 584GEOGRAPHIC ......................................................................................................................................................................................................... 585

COUNTY ....................................................................................................................................................................................................... 585SEGMENT ..................................................................................................................................................................................................... 586BLOCK .......................................................................................................................................................................................................... 587

SAMPLE WEIGHTING AND ESTIMATION VARS ............................................................................................................................................. 588INDEX ...................................................................................................................................................................................................................... 589

APPENDIX A Drug Codes for Open-Ended Questions APPENDIX B Consolidated Drug Other Specify Tables APPENDIX C Tobacco Brand Codes for Open-Ended Questions APPENDIX D Recoded Substance Dependence and Abuse Variable Documentation APPENDIX E Recoded Mental Health APPENDIX F Recoded Depression Variable Documentation APPENDIX G Key Mental Health Variables APPENDIX H Sample SUDAAN®, SAS®, and Stata Code and PUF Estimates

i-1

Introduction to the 2018 National Survey on Drug Use and Health

Introduction

This codebook provides documentation for the 2018 National Survey on Drug Use and Health (NSDUH) public use data file.1 Prior to 2002, the survey was called the National Household Survey on Drug Abuse (NHSDA). NSDUH is sponsored by the Center for Behavioral Health Statistics and Quality (CBHSQ, formerly the Office of Applied Studies) within the Substance Abuse and Mental Health Services Administration (SAMHSA) and is conducted by RTI International, Research Triangle Park, North Carolina.2

For each variable in the 2018 data file, the codebook provides the variable name, a description of the variable, value codes and their meanings, and an unweighted univariate frequency distribution. Most of the variables originated directly as interview items. For a subset of variables created from more than one variable, the source variables and recoding specifications are provided. In addition, case identification, sampling, and data collection variables are included in the file and documented in the codebook. The variable names in this codebook are the variable names used in the dataset.

To a great extent, variable names indicate the meaning of each variable. As much as possible, variables are named consistently across surveys when the content of questions was identical or similar or when the specifications for creating variables were not altered. When a variable does not have the same name across years, this is an indication that differences may exist. In such situations, analysts are advised to examine the codebook documentation in detail before conducting any analysis, particularly those involving the following:

• analyses performed in prior years that are being replicated in the current year,• comparison of data across multiple years, or• use of pooled data from multiple years.

Documentation of the reasons for changing a variable's name can be found in comments above the specific variables and in corresponding appendix files within the codebook (if applicable). Typically, variable changes are documented the year they occur and in the following year if there are comparability issues to consider. However, the most current version of the documentation may not always describe relevant changes from prior years. Therefore, analysts who are doing a multiyear trend analysis are especially advised to examine documentation for all years of interest to identify any change over time that may affect the analysis of interest.3

1 Information on access to NSDUH restricted-use confidential data files is available from the Substance Abuse and Mental Health Data Archive (SAMHDA) at https://datafiles.samhsa.gov/.

2 RTI International is a registered trademark and a trade name of Research Triangle Institute. 3 A variable crosswalk (also referred to as a "measles chart") that identifies which NSDUH variables are

comparable across study years is available on the SAMHDA website at https://datafiles.samhsa.gov/.

https://datafiles.samhsa.gov/https://datafiles.samhsa.gov/

i-2

An analysis of trends is not recommended when using data where the impact of the questionnaire or methodological changes are confounded with true changes in the phenomenon being measured. If the change is considered minor and the likely impact on trend data would be minimal, then comparisons between years may be appropriate.

Analysts also are encouraged to refer to the instrument specifications for the 2018 survey in conjunction with their review of the codebook. The specifications for the 2018 survey provide detailed information about the logic governing how respondents were routed through the questions in the interview and any changes to the instrument relative to the 2017 survey. The 2018 specifications may be found on the SAMHSA website at https://www.samhsa.gov/data/.

Overview of NSDUH

The 2018 NSDUH is the 38th in a series, the primary purpose of which is to measure the prevalence and correlates of substance use and mental health issues in the United States. This survey series provides information about the use of illicit drugs, alcohol, and tobacco among members of the U.S. civilian, noninstitutionalized population aged 12 years old or older. The survey also includes several modules of questions focusing on mental health issues. Surveys have been conducted periodically since 1971, with the most recent ones in 1979, 1982, 1985, 1988, and each year from 1990 through 2018. Currently, public use files are available for surveys from 1979 onward.

Survey Redesign in 1999 and Improvements in 2002

In 1999, the survey underwent a major redesign. The method of data collection was changed from a paper questionnaire administration to a computer-assisted administration. In addition, the sample design was changed from a strictly national design to a state-based sampling plan. These important changes have had a major impact on the data produced from the survey. Because of the differences in methodology and the impact of the new design on data collection, only limited comparisons can be made between data from the redesigned surveys (1999 onward) and data obtained from surveys prior to 1999. Also, because of improvements made to the survey in 2002, the 2018 data in this file should not be compared with data collected in 2001 or earlier to assess changes over time. However, this is not an issue for data from 2002 and data files in subsequent data years, as long as other changes to survey items since 2002 (e.g., changes to question wording) have not affected the comparability of the variables of interest across the surveys.

Prior to 1999, the survey was conducted using paper-and-pencil interviewing (PAPI) methods for an interview lasting about an hour. The PAPI instrumentation consisted of a questionnaire booklet completed by a field interviewer (FI) and a set of individual answer sheets completed by a respondent. All substance use questions and other sensitive questions appeared on the answer sheets, so the FI was unaware of the respondent's answers. Less sensitive questions concerning demographics, occupational status, household size, and composition were asked aloud by the FI and recorded in the questionnaire booklet.

Since 1999, the interview has been carried out using computer-assisted interviewing (CAI) methods. The survey uses a combination of computer-assisted personal interviewing

https://www.samhsa.gov/data/

i-3

(CAPI) conducted by an FI and audio computer-assisted self-interviewing (ACASI). For the most part, questions previously administered by an FI in the PAPI format are now administered by an FI using CAPI. Questions previously administered in the PAPI format using self-administered answer sheets are now administered using ACASI. Use of ACASI is designed to provide respondents with a highly private and confidential means of responding to questions and to increase the level of honest reporting of illicit drug use and other sensitive behaviors.

Partial Questionnaire Redesign in 2015

The NSDUH questionnaire underwent a partial redesign in 2015 to improve the quality of the NSDUH data and to address the changing needs of policymakers and researchers with regard to substance use and mental health issues. Details are provided in the section on Questionnaire Changes for the 2015 NSDUH in the codebook introduction for the 2015 NSDUH public use file on the SAMHDA website at https://datafiles.samhsa.gov/. Details on the 2015 NSDUH questionnaire changes, reasons for the changes, and implications of the changes for NSDUH data users are included in a brief report on these questionnaire changes, in a report on the design changes for the 2014 and 2015 NSDUHs, and in the methodological summary and definitions report for 2015.4,5,6

Where measures existed prior to 2015, these changes may have led to actual or potential breaks in the comparability of data in 2015 with corresponding data from prior years. Where actual or potential breaks in comparability occurred, new variables were created for 2015. Measures not affected by the 2015 partial redesign typically retained the same variable names as in prior years.

The number of allowable characters in a variable name also was expanded for new variables created for 2015. Specifically, the number of allowable characters was expanded from a limit of 8 characters in 2014 and prior years to limits in 2015 of up to 9 characters for edited variables, up to 11 characters for imputed variables, and up to 12 characters for recoded variables. Variables for which names did not change between 2014 and 2015 retained the eight-character limit.

Summary of Information for 2018

The target population for the 2018 survey was the same as has been defined since the 1991 survey: the civilian, noninstitutionalized population of the United States (including civilians living on military bases) who were 12 years of age or older at the time of the survey. Before 1991, the sample was drawn from the household population of the contiguous 48 states. Residents of Alaska and Hawaii were added to the sample population in 1991, as were residents of noninstitutional group quarters (e.g., college dormitories, group homes, civilians dwelling on

4 Center for Behavioral Health Statistics and Quality. (2016). 2015 National Survey on Drug Use and Health: Summary of the effects of the 2015 NSDUH questionnaire redesign: Implications for data users. Retrieved from https://www.samhsa.gov/data/

5 Center for Behavioral Health Statistics and Quality. (2015). National Survey on Drug Use and Health: 2014 and 2015 redesign changes. Retrieved from https://www.samhsa.gov/data/

6 Center for Behavioral Health Statistics and Quality. (2016). 2015 National Survey on Drug Use and Health: Methodological summary and definitions. Retrieved from https://www.samhsa.gov/data/

https://datafiles.samhsa.gov/https://www.samhsa.gov/data/https://www.samhsa.gov/data/https://www.samhsa.gov/data/

i-4

military installations) and persons with no permanent residence (homeless people in shelters and long-term residents of single rooms in hotels). In addition, six special-interest metropolitan statistical areas (MSAs) were oversampled in 1991. The 1992 and 1993 surveys retained the oversampling of the six MSAs and also were designed to provide quarterly as well as annual estimates.

Since 1999, the survey sample has employed a 50-state design with an independent, multistage area probability sample for each of the 50 states and the District of Columbia. For the 1999 through 2013 surveys, the 8 states with the largest populations (which together account for 48 percent of the total U.S. population aged 12 or older according to the 2010 census) were designated as large sample states (California, Florida, Illinois, Michigan, New York, Ohio, Pennsylvania, and Texas) with a target sample size of 3,600, and the remaining 42 states and the District of Columbia had target sample sizes of 900. The 2014 through 2017 sample redesign allowed for a more cost-efficient sample allocation to the largest states, while maintaining sufficient sample sizes in the smaller states to support small area estimation at the state and substate levels. A large reserve sample of area clusters or segments was selected at the time the 2014 through 2017 NSDUH sample was selected, which is being used to field the 2018 through 2022 NSDUHs. Thus, the 2018 through 2022 NSDUH designs simply continue the coordinated design. The 2014 through 2022 NSDUHs were designed to yield 4,560 completed interviews in California; 3,300 completed interviews each in Florida, New York, and Texas; 2,400 completed interviews each in Illinois, Michigan, Ohio, and Pennsylvania; 1,500 completed interviews each in Georgia, New Jersey, North Carolina, and Virginia; 967 completed interviews in Hawaii; and 960 completed interviews in each of the remaining 37 states and the District of Columbia. Consistent with previous designs, the 2014 through 2022 design also oversamples youths aged 12 to 17 and young adults aged 18 to 25. However, the 2014 through 2022 design places more sample in the 26 or older age groups to more accurately estimate drug use and related mental health measures among the aging population. The 2018 sample was allocated to age groups as follows: 25 percent for youths aged 12 to 17, 25 percent for young adults aged 18 to 25, 15 percent for adults aged 26 to 34, 20 percent for adults aged 35 to 49, and 15 percent for adults aged 50 or older.

Important methodological differences since 2002 also have affected the comparability of estimates from 1999 to 2001 with estimates from 2002 and later years. As noted above, the survey's name was changed from the National Household Survey on Drug Abuse (NHSDA) to the National Survey on Drug Use and Health (NSDUH) in 2002. In addition to the survey's name change, each NSDUH respondent since 2002 has been given an incentive of $30. These changes resulted in some improvement in the survey response rate. The changes also affected respondents' reporting of many critical items that are the basis of prevalence measures reported by the survey each year. Further, the data could have been affected by improved data collection quality control procedures introduced in the survey beginning in 2001. In addition, new population data from the 2000 decennial census became available for use in NSDUH's sample weighting procedures, resulting in another discontinuity between estimates since 2002 and those prior to 2002. Where estimates in 2018 are still comparable with those from prior years, analyses of the effects of each of these factors on NSDUH estimates have shown that the 2002 to 2018 data should not be compared with 2001 and earlier data to assess changes over time. For measures comparable across the years being analyzed, however, comparisons may be made between 2002 and surveys in subsequent years, including this one, as noted above.

i-5

A description of the 2018 survey, including more detailed information on sample issues, can be found in several NSDUH documents that discuss methodological issues for 2018 or present some key 2018 findings in tabular, graphical, and report formats. The following 2018 NSDUH documents were released simultaneously on August 20, 2019:

• Center for Behavioral Health Statistics and Quality. (2019). 2018 National Survey on DrugUse and Health: Methodological summary and definitions. Retrieved fromhttps://www.samhsa.gov/data/

• Substance Abuse and Mental Health Services Administration. (2019). Key substance use andmental health indicators in the United States: Results from the 2018 National Survey onDrug Use and Health (HHS Publication No. PEP19-5068, NSDUH Series H-54). Rockville,MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and MentalHealth Services Administration. Retrieved from https://www.samhsa.gov/data/

• Center for Behavioral Health Statistics and Quality. (2019). Results from the 2018 NationalSurvey on Drug Use and Health: Detailed tables. Retrieved fromhttps://www.samhsa.gov/data/

Questionnaire Changes for the 2018 NSDUH

This section describes specific changes to the 2018 NSDUH instrument relative to 2017 that are relevant to variables on the 2018 public use file. As noted previously, names for the corresponding variables were changed in 2018 when it was determined these changes resulted in actual or potential breaks in the comparability of data.

• The logic for identifying respondents who initiated the misuse of prescription pain relievers,tranquilizers, stimulants, or sedatives in the past 30 days was updated to prevent respondentsfrom providing inconsistent answers for their initiation of misuse and their most recentmisuse. This change did not affect the comparability of the data.

• Four new questions for adults were added to the consumption of alcohol section. Thesequestions asked adult respondents whether they thought they ever had a problem with theiruse of alcohol or other drugs or they ever had a problem with their mental health. Adults whothought they had a problem with their alcohol or other drug use were asked whether theyconsidered themselves to be in recovery or to have recovered from the problem with theiralcohol or other drug use. Adults who thought they had a problem with their mental healthwere asked whether they considered themselves to be in recovery or to have recovered fromthe problem with their mental health.

• The section on market information for marijuana that had last been in the interview for the2014 NSDUH was added back to the interview between the consumption of alcohol sectionand the self-administered demographic questions. The content of this section in 2018 was thesame as in the 2014 NSDUH.

The 2018 specifications may be found on the SAMHSA website (for details, see this codebook's Introduction).

https://www.samhsa.gov/data/https://www.samhsa.gov/data/https://www.samhsa.gov/data/

i-6

Additional Historical Changes to NSDUH Variables of Note

This section documents notable changes to NSDUH variables that occurred recently but prior to 2018. Consequently, these earlier changes do not affect the comparability of relevant NSDUH variables in 2018 with variables on data files for NSDUH years following the implementation of these changes.

Extensive changes were made to NSDUH variables in 2015 because of the partial redesign of the 2015 NSDUH questionnaire. These changes are described in the section on Questionnaire Changes for the 2015 NSDUH in the codebook introduction for the 2015 NSDUH public use file on the SAMHDA website at https://datafiles.samhsa.gov/.

Since 2008, a statistical model has been employed to predict whether an adult NSDUH respondent had serious, moderate, mild, or no mental illness based on his or her answers to certain NSDUH questions. For the 2012 NSDUH, this model was revised to produce more accurate estimates. This revised 2012 model was used to produce the mental illness estimates for 2012 to 2018. In addition, comparable estimates were recomputed for 2008 through 2011. For more information on how mental illness estimates were calculated, see this codebook's appendix titled Recoded Mental Health. The mental illness variables included in the 2018 NSDUH data files are based on the revised 2012 model. Note that the mental illness variables (e.g., SMIYR_U, AMIYR_U) are not recommended to be used when analyzing variables for past year suicidal thoughts, past year major depressive episode (MDE), the Kessler-6 (K6) distress scale, or the World Health Organization Disability Assessment Schedule (WHODAS) scale, nor are they recommended to be used when analyzing other variables closely linked with these variables (including past year suicide attempts, suicide plans, medical treatment for suicide attempts, lifetime MDE, serious psychological distress [SPD], or components used in the K6 or WHODAS scales). For more details, see this codebook's appendix titled Recoded Mental Health.

In the 2013 questionnaire and onward, two new response categories were added to the question about race (QD05) in response to changes in U.S. Department of Health and Human Services (HHS) standards of data collection. The response options now include "Guamanian/Chamorro" and "Samoan." To accommodate this change, two new imputation-revised variables were created along with their corresponding imputation indicators: IRRACEGC for Guamanian/Chamorro and IRRACESM for Samoan.7 The change to QD05 also affected the creation of the "Other Pacific Islander" variable IRRACEPI. Prior to 2013, IRRACEPI covered Pacific Islanders who were not Native Hawaiians. Beginning in 2013, the variable name was changed to IRRACPI2 and now covers Pacific Islanders who are not Native Hawaiians, Guamanians/Chamorros, or Samoans. Despite the change to the questionnaire, the levels of the detailed race variable IRNWRACE were not changed. Guamanian/Chamorro and Samoan respondents continued to be treated as Other Pacific Islanders. Therefore, little to no impact is expected on the race/ethnicity variables IRNWRACE, NEWRACE1, NEWRACE2, EXPRACE, and EXPRACE2.

7 For an explanation of imputation-revised variables, see the section on Statistical Imputation in this

codebook introduction.

https://datafiles.samhsa.gov/

i-7

Since 2013, questions also have been included that asked about height and weight. Respondents were asked to report their weight in pounds or kilograms and their height in feet, inches, meters, or centimeters. Edited and recoded variables were created that converted all of the responses from the weight variable into pounds (WTPOUND2) and the height data into inches (HTINCHE2). Using these data, a recoded variable measuring body mass index (BMI2) was included on the public use file, and the documentation can be found in the Recoded Pregnancy and Health Care section of this codebook. Note that this variable has not been used in the detailed tables. The recoded variable called BMI2 was created as a continuous variable based on the following formula: BMI2 = [WTPOUND2 ÷ (HTINCHE2)2] × 703. Unknown response values for the edited weight variable (WTPOUND2) of 9985, 9994, 9997, or 9998 or the edited height variable (HTINCHE2) of 985, 994, 997, or 998 were defined as SAS® missing for the recoded BMI2.8

Strengths and Limitations of NSDUH

NSDUH is the primary nationally representative source of annual estimates of drug use and mental illness among civilian members of the noninstitutionalized population in the United States. Most of the questions in NSDUH are administered with ACASI, which is designed to provide the respondent with a highly private and confidential mode for responding to questions in order to increase the level of honest reporting of illicit drug use and other sensitive behaviors. In addition, the large and dispersed NSDUH sample enables not only state-level estimates, but also estimates for substate areas.

Although NSDUH is useful for many purposes, it has certain limitations. First, the data are based on self-reports of drug use, and their value depends on respondents' truthfulness and memory. Although some experimental studies have established the validity of self-reported data in similar contexts and NSDUH procedures were designed to encourage honesty and recall, some underreporting and overreporting may take place. To aid respondent recall, the prescription drug questions in 2018 allowed respondents to report any use or misuse in the past 12 months of specific related medications (e.g., the tranquilizers Xanax®, Xanax® XR, generic alprazolam, and generic extended-release alprazolam). These self-reports capture information on the use or misuse of prescription drugs that contain a given active ingredient. However, these self-reports are not necessarily accurate for identifying the exact drugs respondents took, especially when respondents identify certain drugs by their brand names (e.g., if a respondent actually took the generic drug alprazolam but reported use or misuse of the brand name tranquilizer Xanax® because of name recognition). For this reason, the public use file includes recoded variables for subtypes of related prescription drugs (e.g., alprazolam products) but with few exceptions does not include variables for individual prescription drugs.

Second, the survey is cross-sectional rather than longitudinal. That is, individuals were interviewed only once and were not followed for additional interviews in subsequent years. Each year's survey, therefore, provides an overview of the prevalence of drug use at a specific point in time rather than a view of how drug use changes over time for specific individuals.

8 The formula for calculating BMI for use in BMI2 is from the Centers for Disease Control and Prevention

(CDC) website: https://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html?s_cid=tw_ob064.

https://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html?s_cid=tw_ob064

i-8

Third, because the target population of the survey is defined as the civilian, noninstitutionalized population of the United States, a small proportion (approximately 3 percent) of the population is excluded.9 The subpopulations excluded are members of the active-duty military and individuals in institutional group quarters (e.g., hospitals, prisons, nursing homes, and treatment centers). If the drug use of these groups differs from that of the civilian, noninstitutionalized population, NSDUH may provide slightly inaccurate estimates of substance use and mental health in the total population. This may be particularly true for prevalence estimates for less commonly used drugs, such as heroin.

Survey Methodology

Like the 1999 to 2017 surveys, the 2018 survey was conducted using CAI methods. This survey also allows for improved state estimates based on minimum sample sizes per state. The target sample size of 67,507 allows SAMHSA to continue reporting adequately precise demographic subgroup estimates at the national level without needing to oversample specially targeted demographics, as was required in the past. The achieved sample size, or completed interviews, for the 2018 survey was 67,791 individuals.

A coordinated sample design was developed for the 2014 through 2017 NSDUHs. A large reserve sample of area clusters or segments was selected at the time the 2014 through 2017 NSDUH sample was selected. This reserve sample is being used to field the 2018 through 2022 NSDUHs. Thus, the 2018 through 2022 NSDUH designs simply continue the coordinated design. The coordinated design facilitates a 50 percent overlap in third-stage units (area segments [see below]) between each 2 successive years from 2014 through 2022.10 This design was intended to increase the precision of estimates in year-to-year trend analyses because of the expected positive correlation resulting from the overlapping sample between successive survey years.

The 2018 design allows for computation of estimates by state in all 50 states plus the District of Columbia. States may therefore be viewed as the first level of stratification and as a reporting variable. Compared with previous sample designs, the 2014 through 2022 sample design moves from two to essentially five state sample size groups (including Hawaii with the remaining states and the District of Columbia). The 2014 through 2022 surveys have a sample designed to yield 4,560 completed interviews in California; 3,300 completed interviews each in Florida, New York, and Texas; 2,400 completed interviews each in Illinois, Michigan, Ohio, and Pennsylvania; 1,500 completed interviews each in Georgia, New Jersey, North Carolina, and Virginia; 967 completed interviews in Hawaii; and 960 completed interviews in each of the remaining 37 states and the District of Columbia—for a total national target sample size of 67,507. The sample is selected from 6,000 area segments that vary in size according to state. The change in the state sample allocation in 2014 was driven by the need to increase the sample

9 Information from the 2010 census suggests that the civilian, noninstitutionalized population includes at

least 97 percent of the total U.S. population. See the following reference: Lofquist, D., Lugaila, T., O'Connell, M., & Feliz, S. (2012, April). Households and families: 2010 (C2010BR-14, 2010 Census Briefs). Retrieved from https://www.census.gov/prod/cen2010/briefs/c2010br-14.pdf

10 In segments used in 2 successive years, only addresses not sampled in the first year may be included in the second year's sample.

https://www.census.gov/prod/cen2010/briefs/c2010br-14.pdf

i-9

in the original 43 small states (to improve the precision of state and substate estimates in these states) while moving closer to a proportional allocation in the larger states.

Stratification and Selection of Primary, Secondary, and Tertiary Sampling Units (Census Tracts, Census Block Groups, and Area Segments)

Within each state, sampling strata called state sampling regions (SSRs) were formed. Based on a composite size measure, states were partitioned geographically into roughly equally sized regions. In other words, regions were formed such that each area within a state yielded, in expectation, roughly the same number of interviews during each data collection period. The partitioning divided the United States into a total of 750 SSRs, resulting from 36 SSRs in California; 30 SSRs each in Florida, New York, and Texas; 24 SSRs each in Illinois, Michigan, Ohio, and Pennsylvania; 15 SSRs each in Georgia, New Jersey, North Carolina, and Virginia; and 12 SSRs each in the remaining 38 states and the District of Columbia.

Similar to the 2005 through 2013 surveys, the first stage of selection for the 2014 through 2022 NSDUHs was census tracts. The first stage of selection began with the construction of an area sample frame that contained one record for each census tract in the United States. If necessary, census tracts were aggregated within SSRs until each tract11 met the minimum dwelling unit12 (DU) requirement. In California, Florida, Georgia, Illinois, Michigan, New Jersey, New York, North Carolina, Ohio, Pennsylvania, Texas, and Virginia, this minimum size requirement was 250 DUs in urban areas and 200 DUs in rural areas.13 In the remaining states and the District of Columbia, the minimum requirement was 150 DUs in urban areas and 100 DUs in rural areas. These census tracts served as the primary sampling units (PSUs) for the coordinated 9-year sample.

Before selecting census tracts, additional implicit stratification was achieved by sorting the first-stage sampling units by a CBSA/SES14 (core-based statistical area/socioeconomic status) indicator15 and by the percentage of the population who are non-Hispanic and white. From this well-ordered sample frame, 48 census tracts per SSR were selected with probabilities proportionate to a composite size measure and with minimum replacement.

For the second stage of selection, adjacent census block groups were collapsed as needed within selected census tracts. Compared with years prior to 2014, the selection of census block

11 For the remainder of the discussion, first-stage sampling units are referred to as "census tracts" even though each first-stage sampling unit contains one or more census tracts.

12 DU counts were obtained from the 2010 decennial census data supplemented with revised population counts from Claritas, which is a market research firm headquartered in Ithaca, New York (see https://www.claritas.com/ ).

13 The basis for the differing minimum DU requirement in urban and rural areas is that it is more difficult to meet the requirement in rural areas; 100 DUs are sufficient to support one field test and two main study samples in the smaller states, and 200 DUs are sufficient to support three samples in the larger sample states.

14 CBSAs include metropolitan and micropolitan statistical areas, as defined in the following reference: Office of Management and Budget. (2009, December 1). OMB Bulletin No. 10-02: Update of statistical area definitions and guidance on their uses. Washington, DC: The White House.

15 The CBSA/SES indicator was defined using 2006-2010 American Community Survey (ACS) estimates, 2010 census data, and the December 2009 CBSA definition. Four categories are defined as follows: (1) CBSA/low SES, (2) CBSA/high SES, (3) non-CBSA/low SES, and (4) non-CBSA/high SES.

https://www.claritas.com/

i-10

group is an additional stage of selection that was added to facilitate possible transitioning to an address-based sample (ABS) design in the future. The block groups were required to have the same minimum number of DUs as the census tracts from which they were selected (150 or 250 in urban areas and 100 or 200 in rural areas, according to state). The resulting block groups were then sorted in the order in which they were formed, and one census block group16 was selected per selected census tract with probability proportionate to a composite size measure.

Because census block groups generally exceed the minimum DU requirement, one smaller geographic area was selected within each sampled census block group. For this third stage of sampling, each selected census block group was partitioned into small geographic areas composed of adjacent census blocks. These geographic clusters of blocks are referred to as segments and are the tertiary sampling units (TSUs) for the coordinated sample design. A sample DU in NSDUH refers to either a housing unit or a group quarters listing unit, such as a dormitory room or a shelter bed. To support the overlapping sample design and any special supplemental samples or field tests SAMHSA might wish to conduct, segments were formed to contain a minimum of 150 or 250 DUs in urban areas and 100 or 200 DUs in rural areas, according to state.

One segment was selected within each sampled census block group with probability proportionate to size. The 48 selected segments then were randomly assigned to a survey year and quarter of data collection.

Selection of Dwelling Units

The primary objective of the fourth stage of sample selection (listing units) was to select the minimum number of DUs needed in each segment to meet the targeted sample sizes for all age groups. For the 2014 through 2022 NSDUHs, each state sample was allocated to age groups as follows: 25 percent for youths aged 12 to 17, 25 percent for young adults aged 18 to 25, 15 percent for adults aged 26 to 34, 20 percent for adults aged 35 to 49, and 15 percent for adults aged 50 or older. In the 2005 through 2013 NSDUHs, the sample was allocated equally across the 12 to 17, 18 to 25, and 26 or older age groups. The 2014 through 2022 design places more sample in the 26 or older age groups to estimate drug use and related mental health measures more accurately among the aging population. The size measures used in selecting the area segments were coordinated with the DU and person selection process so that a nearly self-weighting sample could be achieved in each of the five age groups. Departures from the self-weighting objective occurred for several reasons, including the following: (a) advance projections on the number of DUs did not accurately reflect the current housing inventory; (b) maximum DU sample sizes were preset to control the interviewer workload and to allow unused addresses to be available for the next year's survey; and (c) the person selection probabilities were constrained so that no more than two individuals could be selected per DU. An iterative sample allocation process was followed to adjust for these additional constraints. In addition, the DU sample allocation in each area segment was adjusted to allow for DU eligibility, for screening nonresponse, and for person nonresponse.

16 For the remainder of the discussion, second-stage sampling units are referred to as "census block groups"

even though each second-stage sampling unit contains one or more census block groups.

i-11

In advance of the survey period, specially trained listers had visited each area segment and listed all addresses for housing units and eligible group quarters units in a prescribed order. Systematic sampling was used to select the allocated sample of addresses from each segment.17

Selection of Individuals

During each quarterly survey, FIs visited each sample address to determine DU eligibility, to list all eligible individuals at the address, to select the sample of individuals to be interviewed (if any), and to conduct interviews. Unlike the 2005 through 2013 NSDUHs, the "half-open" interval (HOI) rule was not implemented in the 2014 through 2018 NSDUHs. This special procedure identified any new (since the time of listing) housing units or any DUs missed during the advance listing process. Any new or missed DUs following immediately after a sample DU and up to, but not including, the next initially listed address in the prescribed order of listing also were included in the sample. Eliminating the HOI rule in 2014 decreased the burden on interviewers and simplified training and the screening process. This decrease in burden outweighed the small amount of coverage afforded by the HOI rule. Because the majority of missed DUs are found on the premises of sampled DUs, the 2014 through 2018 NSDUHs had in place a procedure for checking for and adding missed DUs on the premises of sampled DUs. During the screening interview, FIs asked the screening respondent about other units on the property of the selected DU. If missing from the original list, these DUs also were included in the sample.

The FIs used a handheld computer to record the results of the DU screening process and to select the sample of respondents. They recorded the results of each visit, the final eligibility status of the DU, and information on new and missed DUs. If a sample address was an eligible occupied DU, the FIs also conducted a screening interview to identify and roster all survey-eligible individuals residing at the address. When the roster was complete, the computer was programmed to select the sample of individuals to be interviewed using parameters specified for that area segment and a random number specified for that address.

Data collection progress was monitored during each quarterly survey by state. Small reserve samples were held back each quarter so that the assigned sample size could be adjusted if necessary during the course of data collection.

17 This had the effect of creating noncompact clusters (selection from a list), which differ from compact

clusters in that not all units within the cluster are included in the sample. Although compact cluster designs are less costly and more stable, a noncompact cluster design was used because it provides for greater heterogeneity of dwellings within the sample. Also, social interaction (contagion) among neighboring dwellings is sometimes introduced with compact clusters. See the following reference: Kish, L. (1965). Survey sampling. New York, NY: John Wiley.

i-12

Sample Design Variables

The sample for the 2018 NSDUH was selected using a multistage, deeply stratified sample design. The variables on the full restricted-use analytic data file18 represent each stage of sample selection (note that, to ensure the confidentiality of survey respondents, the sample design variables described in this section are not included on the public use file):

Stage 1: The 2018 NSDUH design includes a sample from each of the 50 states plus the District of Columbia. SSRs were formed within each state based on composite size measures, roughly geographically partitioning the state into equally sized regions. A sample of 48 census tracts was selected within each SSR. Census tracts are considered the PSU and can be identified using the STATE, SSREGION, and SEGID19 variables on the full restricted-use analytic data file. Only eight census tracts per SSR were used for the 2018 sample.

Stage 2: For the second stage of selection, adjacent census block groups were aggregated within selected census tracts as necessary to form the second-stage sampling units. One census block group was selected per sampled census tract. Within each SSR, 48 census block groups were selected. Census block groups can be identified using the STATE, SSREGION, and SEGID variables on the full restricted-use analytic data file. Eight census block groups per SSR, one from each sampled census tract, were used for the 2018 sample.

Stage 3: The third stage of sampling consisted of partitioning the selected census block groups into smaller geographic areas, called "segments." Segments are defined by joining contiguous census blocks within each selected census block group and are similar to the units selected at the second stage of selection for the 2005 through 2013 surveys. Segments can be identified on the full restricted-use analytic data file using the SEGID variable.

Stage 4: After census tracts, census block groups, and segments were selected, the fourth stage of selection consisted of selecting DUs within each segment. The DU selection rate was based on the state to which a particular segment belonged. State classification was utilized for computation of eligibility, screener and interview response rates, and expected person yield per DU. On the full restricted-use analytic data file, state classifications can be identified with the STATE variable, and DUs can be identified with the ENCCASE variable.

18 NSDUH's Restricted-use Data Analysis System (R-DAS) data files are available online for most years at

https://datafiles.samhsa.gov/. R-DAS is an online analytic system that allows analysts to produce cross-tabulations using restricted-use NSDUH data files. Restricted-use microdata are not accessible to analysts, but output from the analyses is available as long as the output does not violate any of the disclosure limitation rules that determine what output may be displayed. Because of NSDUH's questionnaire redesign in 2015 and subsequent breaks in trends for a number of variables, an R-DAS file was not produced for the 2015 NSDUH data. R-DAS files have been created, however, for aggregated 2015-2016 and 2016-2017 NSDUH data, as well as for earlier aggregated 2-year pairs of data.

19 One segment was selected from each sampled census block group within each sampled census tract, so the census tract, census block group, and segment can be identified by the SEGID variable.

https://datafiles.samhsa.gov/

i-13

Stage 5: At the last stage of selection, individuals were selected within screened DUs based on the age group composition of the DU residents. The person-level variable used to determine the selection included AGE. The full restricted-use analytic data file contains one record representing each responding selected person from stage 5 (67,791 individuals). To protect the confidentiality of these respondents, the full analytic file was treated using a statistical disclosure limitation method while ensuring that the data continue to be representative of civilian members of the noninstitutionalized population in the United States. The resulting public use data file contains 56,313 records.

A more detailed explanation of the sample design and sample selection procedures at each stage of the design appears in the 2018 NSDUH Methodological Resource Book, which will be available in 2020.20

Data Collection and Response Rates

The fieldwork for the 2018 NSDUH was directed by RTI staff members. RTI maintained a field staff of approximately 650 FIs to collect the data.

A total final sample of 67,791 interviews was obtained for the 2018 survey. Strategies for ensuring high rates of participation resulted in a weighted screening response rate of 73.30 percent and a weighted interview response rate for the NSDUH of 66.56 percent; the overall response rate was 48.79 percent for people aged 12 or older.

Throughout the course of the study, respondent anonymity and the privacy of responses were protected by separating identifying information from survey responses. Respondents were assured that their identities and responses would be handled in strict compliance with federal law. As discussed above, the questionnaire itself and the interviewing procedures were designed to enhance the privacy of responses, especially during segments of the interview in which questions of a sensitive nature were posed. Answers to sensitive questions were gathered using ACASI. During the ACASI portions of the interview, respondents listened to prerecorded questions through headphones and entered their responses directly into a NSDUH laptop computer without FIs knowing how they were answering. At the conclusion of the ACASI section, the interview returned to the CAPI mode with the FI asking the respondent questions and entering the responses into the laptop. Respondents who completed a full interview were each given $30 in cash as a token of appreciation for their time.

A more detailed explanation of NSDUH's data collection procedures appears in the 2018 NSDUH Methodological Resource Book, which will be available in 2020.21

20 Center for Behavioral Health Statistics and Quality. (in press). 2018 National Survey on Drug Use and

Health: Methodological resource book. Rockville, MD: Substance Abuse and Mental Health Services Administration.

21 See the reference in footnote 20.

i-14

Sample Weights

The estimates yielded by NSDUH are based on sample survey data rather than on complete data for the entire population. This means that the data must be weighted to obtain unbiased estimates for survey outcomes in the population represented by the 2018 NSDUH. The "final analysis weight" of the ith respondent, say wi, can be interpreted as the number of sampling units in the NSDUH target population represented by the ith respondent. The sum of the weights over all respondents is used to estimate the size of the total target population:

i estimated size of target population,i

w =∑

where the summation is over all respondents in the 2018 NSDUH.

Similar to the 2017 NSDUH, three sets of analysis weights at the person level, questionnaire dwelling unit (QDU) level, and person pair level were developed for the 2018 NSDUH. The person-level, QDU-level, and person pair-level analysis weights shared the same first 11 weight components at the screening dwelling unit (SDU) level. In addition to the 11 common weight components, QDU-level and person pair-level analysis weights had several specific weight components, and the final weights are the product of all of the weight components. As in the 2017 NSDUH, all of the adults in the 2018 NSDUH sample received the WHODAS questions. Therefore, there was no need to have a separate adult mental health weight in the 2018 NSDUH because the person-level analysis weight could be used to produce the adult mental health estimates.

The person-level analysis weights (ANALWT_C) are the product of 16 weight components from the analytic file, and two additional weight calibration adjustments done for the public use file.22 Each weight component accounts for either a selection probability at a selection stage or an adjustment factor adjusting for nonresponse, coverage, or extreme weights. The sum of the weight over all respondents on the data file represents an estimate of the total number of individuals in the target population. In view of the use of weights as expansion factors in forming estimates, the weight can be interpreted as the total number of individuals in the target population that each record on the file represents. For variance estimation, suitable software, such as SUDAAN®, should be used to take the sample design into account.23 Similar to the 2017 NSDUH, the 2018 NSDUH used 2010 census-based population estimates in the poststratification adjustment.

Details of the weight components and the sample weighting procedures appear in the 2018 NSDUH Methodological Resource Book, which will be available in 2020.24

22 For details, see the section on Public Use File Weight Calibration. 23 In SUDAAN, the sample design is specified using the NEST statement. See the following reference:

RTI International. (2012). SUDAAN®, Release 11.0 [computer software]. Research Triangle Park, NC: Author. 24 See footnote 20.

i-15

Organization of the Data File

The file described here is made available as an ASCII file with 2,691 variables and 56,313 observations. Three program files are made available to read the ASCII file into SAS, SPSS, or Stata. The file also is made available as a SAS transport (CPORT) file, SPSS system file, Stata system file, and ASCII tab-delimited file. All of the data and program files are available from SAMHDA at https://datafiles.samhsa.gov/.

The overall organization of the file is shown in the Table of Contents of this document. The first portion of this codebook consists of documentation for edited variables from drug modules, followed by documentation for imputation-revised and selected recoded versions of these drug variables. Documentation for edited data from the second set of self-administered modules and demographic questions are in later sections; each of these later sections is immediately followed by sections corresponding to imputation revised and recoded variables, where applicable. Where imputed or recoded variables are provided, users are encouraged to use these variables (rather than raw or edited variables) to produce estimates from the interview.

Most edited variables contain missing data (e.g., the edited variable CIGREC for the most recent use of cigarettes); see the section on Standard Code Conventions for a description of the codes given to different types of missing data for edited variables. For each edited variable, the number of observations assigned a given missing data code is shown in the entry for that variable.

Imputation-revised variables, as well as selected recoded versions of these variables, are included for selected demographic variables, drug use variables, and certain substance use disorder (SUD) variables. As noted previously, these imputed and recoded variables are in separate sections following the edited variables. In particular, the recoded drug use variables include indicators for lifetime, past year, and past month substance use. The imputation-revised drug use variables served as the starting point for the recoded drug use variables. Imputed variables for SUDs for a subset of illicit drugs also are included in 2018 and are documented in the Imputed Substance Dependence and Abuse and Recoded Substance Dependence and Abuse sections of the codebook. Imputation-revised demographic variables also are documented toward the end of the codebook. Missing values for all imputation-revised variables have been imputed using the statistical imputation procedures described in the Statistical Imputation section of this introduction. Imputation indicators are provided for each variable so that users may easily determine whether an observation contains data from the questionnaire or an imputed value.

Following the sections for substance use variables, edited variables from additional self-administered modules comprise the next major section in the codebook. Intermixed within this major section of the codebook are sections of recoded variables. The edited and/or imputed variables are used as source variables for the recoded variables. For example, edited variables from the section of the interview pertaining to the receipt of treatment for the use of alcohol or illicit drugs (Drug Treatment section in this codebook) were used to create recoded summary measures for the receipt of substance use treatment. Unlike variables in the drug sections, however, imputed variables were not created for many variables from this second section of the interview, except where noted. Consequently, these recoded variables may still have missing

https://datafiles.samhsa.gov/

i-16

values. The missing data codes contained in the source variables and defined in the Standard Code Conventions subsection within the Logical Editing section are recoded to the standard missing code (.) for recoded variables. It is intended that cases containing these missing codes be excluded from an analysis. Note that in some situations where recoded variables are created from edited variables, missing values are set to a value of 0 instead of to the standard code for missing data (.)

Within the recoded sections of the codebook, detailed information is provided in the variable documentation about how levels of source variables are used to define recoded variables. Note that for recoded variables that include other-specify data or logically assigned data (as defined in the Standard Code Conventions subsection), values may be listed in the documentation that do not exist for the current survey year. Because of fluctuations in the data, source variable values may not exist consistently across years. However, to aid in cross-year analyses, all possible source variable values have been retained in the documentation. For example, a recoded variable may document a "yes" using both source variable values of "1 = Yes" and "3 = Logically Assigned Yes," even if the 3 is not applicable for the current survey year. To alert users, notes have been placed at the top of recoded sections containing variables that include source values even if they do not apply to the current survey year.

In many instances, the codebook itself also indicates in parentheses the question names used in the CAI instrument that were most relevant for creating an edited variable. As much as possible, the codebook shows key source variables used in creating edited, imputed, or recoded variables. In particular, an important feature of the transition to CAI in 1999 was that respondents could be routed to different versions of a question based on prior information from the interview. As much as possible, the codebook shows key source variables used in creating edited, imputed, or recoded variables. For example, respondents who initiated use of a drug within 1 year of their current age were asked more detailed questions about the year and month in which they first used that drug. Depending on their age, date of birth, and interview date, respondents could be routed to one of three possible questions to identify the year in which they first used a drug (e.g., CG04a, CG04b, or CG04c for cigarettes). Similarly, respondents could be routed to one of two possible questions to determine the month in which they first used a drug (e.g., CG04c and CG04d for cigarettes). To facilitate analyses, therefore, responses from these multiple year-of-first-use and month-of-first-use questions were combined into one single year-of-first-use variable and one single month-of-first-use variable for each drug, as shown for the variables CIGYFU and CIGMFU. Thus, the codebook documentation shows that CIGYFU was based on data from the source variables CG04a through CG04c, and CIGMFU was based on data from the source variables CG04c and CG04d. For the most complete information about the logic for asking questions in the interview, however, data file users should refer to the 2018 questionnaire.

Usable Cases

A key step in the data processing procedures established the minimum item response requirements in order for cases to be retained for weighting and further analysis (i.e., "usable" cases). These procedures were designed to eliminate cases with unacceptable levels of item nonresponse (i.e., missing data), thereby retaining cases with lower levels of missing data and reducing the amount of statistical imputation needed for any given record.

i-17

The usable case criteria established for NSDUH were based on the completeness of information respondents provided about their lifetime use or nonuse of different substances. In NSDUH, respondents were asked more detailed questions about different substances only if they reported lifetime use of that substance (or lifetime use of one or more drugs within a broader category, such as hallucinogens) on an initial "gate" question; these questions are subsequently referred to as gate questions because a report of use opens the "gate" to the administration of additional questions, and other reports that do not explicitly indicate use close the "gate."25 Consequently, whether a NSDUH respondent was a user or nonuser of the substances of interest could be readily determined by reviewing the respondent's answers to the gate question on lifetime use of that substance (or category of substances).

For a NSDUH record to be considered usable, both of the following requirements must be met:

1. The lifetime cigarette question (i.e., "Have you ever smoked part or all of a cigarette?") had to have been answered as "yes" or "no." This requirement was set so that lifetime use or nonuse would be fully defined for at least one substance. Consequently, data about lifetime use or nonuse of cigarettes could be used in subsequent statistical imputations for other substances where lifetime use or nonuse was undefined.

2. Responses to questions on at least nine (9) of the following additional substances had to contain information about use or nonuse: (a) smokeless tobacco, (b) cigars, (c) alcohol, (d) marijuana, (e) cocaine (in any form), (f) heroin, (g) hallucinogens, (h) inhalants, (i) methamphetamine, (j) pain relievers, (k) tranquilizers, (l) stimulants (i.e., independent of methamphetamine), and (m) sedatives. Crack cocaine was not included in the usable case rule because the logic for asking about crack cocaine was dependent upon the respondent having answered the lifetime cocaine question as "yes." Although the CAI instrument also asked about pipe tobacco, this was not included in the usable case rule because there was only one other question about pipe tobacco in addition to the lifetime pipe tobacco use question.26,27

The usability criterion for smokeless tobacco through heroin was that lifetime use or nonuse must be determined. This same usability criterion was applied to the module for methamphetamine beginning in 2015. For the "multiple gate" modules for hallucinogens and

25 In all substance use modules except those pertaining to hallucinogens, inhalants, pain relievers,

tranquilizers, stimulants, and sedatives, the logic for asking more detailed questions about use of that drug was based on the answer to a single "yes/no" question (e.g., "Have you ever, even once, used marijuana or hashish?").

26 For a more detailed discussion of the development of the usable case requirements for the transition to CAI in the 1999 survey, see the following reference: Kroutil, L., & Myers, L. (2002). Development of editing rules for CAI substance use data. In J. Gfroerer, J. Eyerman, & J. Chromy (Eds.), Redesigning an ongoing national household survey: Methodological issues (HHS Publication No. SMA 03-3768, pp. 85-109). Rockville, MD: Substance Abuse and Mental Health Services Administration, Office of Applied Studies.

27 For more information about the development of the usable case requirements for the redesigned NSDUH questionnaire, see the following reference: Center for Behavioral Health Statistics and Quality. (2014). National Survey on Drug Use and Health: 2012 Questionnaire field test final report. Retrieved from https://www.samhsa.gov/data/

https://www.samhsa.gov/data/

i-18

inhalants, at least one gate question in the series for that module was required to have an answer of "yes" or "no."28 Any of the following met the usability criteria for prescription drugs:

• past year use of at least one specific prescription drug in a category (e.g., pain relievers) was reported, or

• lifetime use or nonuse of any prescription drug in the category was reported, or • past year nonuse of all specific prescription drugs was reported, regardless of whether

lifetime use or nonuse can be determined.

The interview also included follow-up probes for respondents who initially refused to answer a gate question or an entire series of gate questions. Follow-up probes were included in the following modules that were relevant to the usable case rule: cigarettes, smokeless tobacco, cigars, alcohol, marijuana, cocaine, heroin, specific hallucinogens (i.e., LSD, PCP, Ecstasy/"Molly" [MDMA]), any use of inhalants, and methamphetamine. If respondents changed their initial refusal to a response of "yes" or "no," they were considered to have provided usable data to that drug's gate information.

In addition, the CAI program terminated interviews during the initial demographics questions if respondents were ineligible for the survey (i.e., under age 12 or on active duty in the U.S. military). These ineligible cases by definition did not meet the usable case criteria because the interviews were terminated before the respondents were asked the first cigarette use question.

Logical Editing

For selected key variables, response data were reviewed to identify and address inconsistent data among related variables or to replace missing data with nonmissing values. The routing logic in the CAI instrument reduced the opportunities for respondents to give inconsistent answers by skipping respondents past questions that did not apply to them. The occurrence of inconsistent data was reduced further through the use of consistency checks built into the CAI program that prompted respondents to resolve inconsistencies between related items. Nevertheless, there still were limited situations in which respondents could answer one question in a manner that was inconsistent with their answer to a previous question.

Logical editing was the first step in processing many of the variables on the file. This procedure used data within a respondent's record to identify and address inconsistencies among related variables within a given module of the interview (e.g., within the hallucinogens section). As part of this procedure, variables were identified that had been legitimately skipped because the condition(s) for asking the questions did not apply.

28 For hallucinogens and inhalants, the logic for asking more detailed questions about use of that category

of drugs was based on respondents' answers to multiple "yes/no" questions about the lifetime use or nonuse of specific drugs within that category (e.g., lifetime use or nonuse of the following specific hallucinogens: lysergic acid diethylamide (LSD); phencyclidine (PCP); peyote; mescaline; psilocybin/mushrooms; Ecstasy or "Molly" [3,4-methylenedioxymethamphetamine (MDMA)]; ketamine; DMT, AMT, or "Foxy" (i.e., dimethyltryptamine, alpha-methyltryptamine, and 5-MeO-DIPT [5-methoxy-di-isopropyltryptamine], respectively); Salvia divinorum; or "any other" hallucinogen).

i-19

As a general principle, responses from one module (e.g., hallucinogens) were not used to edit variables in another module (e.g., inhalants).29 For this reason, data in one module may not be completely consistent with data in other modules. Subsequent discussion about editing in this section focuses on key substance use variables.

Editing Procedure for Substance Use Variables Other than Prescription Drugs

This section provides an overview of editing procedures for variables pertaining to substances other than prescription drugs. Editing procedures for prescription drugs are described separately in the following section because the questionnaire is structured differently for the prescription drug questions.

In sections of the interview for tobacco, alcohol, marijuana, cocaine (including crack cocaine), heroin, hallucinogens, inhalants, and methamphetamine, logical editing and processing of substance use variables first involved identifying whether respondents had ever used or never used the substance of interest. Determining whether or not respondents had ever used a given substance (or any substance within a broader category, such as hallucinogens) included identifying situations in which respondents initially refused to answer a question about their lifetime use of a substance but then changed their answer to "yes" or "no" on follow-up. If respondents did not provide sufficient information about their use or nonuse of a particular substance (or category of substances, such as hallucinogens), their final status was assigned through statistical imputation procedures. If values pertaining to lifetime use or nonuse of a given substance were changed through the editing procedures, this editing was indicated through special codes that indicated that a response was logically inferred; documentation for these codes includes the phrase "LOGICALLY ASSIGNED." For example, if respondents did not report lifetime use of LSD (a specific hallucinogen) in question LS01a, but they specified use of it as some other hallucinogen they used, then they were logically inferred to be lifetime users of LSD.

After lifetime use or nonuse of a given substance had been determined as part of the editing procedures, edits were implemented for the variables that established when respondents last used a substance of interest. These edited recency-of-use variables were the precursors for the final, imputed measures that established the prevalence of use in the past 30 days, past 12 months, and lifetime.

The interview included follow-up probes for respondents who were lifetime users of a given substance but did not know or refused to report when they last used it. Respondents who initially did not know when they last used a substance were asked to give their "best guess" of when they last used it. Respondents who initially refused to report when they last used a substance were asked to reconsider answering the question. If respondents changed their initial answer of "don't know" or "refused" in response to these probes, the editing procedures incorporated data from these probes into the final, edited recency-of-use variables. For example, if respondents initially refused to report when they last used a substance but then reported in the

29 One exception to the principle of not editing across modules involved situations in which responses in

one module governed whether respondents were asked questions in another module. For example, if respondents reported never using heroin but they received substance abuse treatment in the past 12 months, they were not asked questions in the drug treatment module about current or past year treatment for heroin. Consequently, codes could be assigned to indicate the respondents skipped out of the heroin treatment questions because they had never used it.

i-20

follow-up probe that they last used it more than 30 days ago but within the past 12 months, their edited recency indicated use in that period, in the absence of any information that was inconsistent with what they reported in the probe (see below).

For substances other than prescription drugs, situations were identified and flagged in which there were inconsistencie